Dustbins Digitalisation and Garbage Collection Route Optimisation
This so called “BINternet” focuses on a smart collection method to optimize cities’ waste management practices. It suggests building a real-time data tracking system and then solve an optimised collection road for garbage truck based on these data. By digitalising and optimising the dustbin management, cities can have a better control over emissions and costs.
It is always a rarity to look beneath the surface and focus on the meaning hidden inside. Through an apple, we discovered the gravity; through the gender pay gap, we fought against gender inequality worldwide; through the price increases, we smelled inflations and predicted financial crises. Likewise, when we talk about the garbage, can we realise that what we are talking about is an energy, but which has been disposed wrongly.
Naturally, a question follows – what should be the right way to tackle the waste? In this proposal, the waste management, we will continuously pay attention to the smart integration into the existing waste management solutions, including collection, storage, and re-use techniques.
At a start of smart waste management concept, , this proposal will cover on a new collection method, which will optimise collection routes for garbage trucks by digitalising the real-time volumes of the dustbins. All these data will be kept and analysed for both location and volume. This will help the garbage transportation centre to create a reliable database, just as the Internet, all input data will be linked together for further analysis and results can be presented, we can call this dustbins database “the BINternet”.
To finally select the optimised route, there are two steps to be followed; the digitalisation, and the optimisation. In the following sections, we will discuss these two steps.
What we are expecting is the optimised route for waste collection. To programme such an optimisation route selection system, we need to analyse incoming data; to get the data, we need to monitor dustbins; to monitor bins, we need to install special tracking sensors. This special sensor is the one who plays an important role for digitalisation process.
However, what are the types of data this sensor should collect? The answer can be diverse for different methods and different requirements, however, here in this proposal, we choose two types of data: the height of the bin, and the location of the latitude and longitude of the bin.
As a usual practice for garbage collection, the garbage truck reaches to all bins inside his assigned area. However, it is usually the case that some dustbins located in relatively rural areas in the city, are not even half-full when it is collected. In other words, the collecting frequency for the dustbins in the downtown and the rural area are quite different. Therefore, it will be more practical if the truck only goes by the full bins and collect them.
We calculate the empty ratio (ER) as the collecting criteria, where
The ER data is monitored by a sensor installed on the bin, and this data is collected real-time. If a bin’s ER is over 80% when the truck is about to leave to collect, this bin is being selected. When the driver is leaving, the app installed in the truck will run automatically to select the optimised route, considering only the selected bins and the rea-time traffic.
Besides the ER data, there is also another data need to be collected and considered, the location. There are lots of methods to measure the concrete location, for instance, the latitude and longitude, and GIS based applications. In this proposal, we choose GIS app and its shortest path application. Using it, we could easily locate each bin and draw them on a map.
To summarize, we locate the bins separated in the city and draw them on a 2D map, while based on real-time empty ratio, we give each bin a monitored data. By doing so, we digitalise the dustbins data, build a database, and transfer a 3D to 2D optimisation problem.
After tracking data and building a database, we can start to analyse the optimised collection route. To finish this, we can choose either Dynamic Programming (DP) or the Shortest Path Problem (SPP). We will discuss them respectively.
Dynamic programming, as a basic algorithm, is a technique which can be used to solve problems in which a sequence of decisions must be made over time. It can also sometimes be used to solve problems which do not involve such a sequence of decisions, but which can be expressed as if they did involve such a sequence. Before giving a formal function of how to use DP for smart waste collection, we first consider a simple example to gain a better understanding.
The Smith family, who live at Kendal in Cumbria, have two children aged 10 and 12 years old. The family have regularly watched a recent television series based around the legend of Robin Hood, a famous English folk hero. The children have become very interested in Robin Hood and read several books about him and his band.
They lived in Sherwood Forest and the legend says that they made their home in a large hollow oak tree in the forest, called the Major Oak. This tree still exists and the children have asked their parents to take them to see it. The parents have agreed to do this the very next Sunday. The Major Oak is situated in the remaining part of Sherwood Forest near a town called Edwinstowe, in Nottinghamshire.
The father has looked at his road maps to work out how to get there. He doesn’t like driving on motorways and has found that there are several alternative ways to get there. He has the choice of going through various towns as shown on the map below. The towns are all labelled by their initial letter and the numbers give the distances in miles between the different towns.
The father expects to drive at the same average speed along any road. Since he wants to do the journey there and back in the same day he wants to find the minimum
distance he should travel.
Starting from Kendal he has a choice of 2 roads, to go to town R or S. When he is at either town R or S, there is a choice of 3 roads to towns W, L or H. There is then a choice of two roads from each of these 3 towns to either town D or B. Finally, from D or B he goes straight to Edwinstowe. Thus, the total number of alternative different routes from which he should choose is 2x3x2 = 12. The distances along each route are as follows.
K R W D E = 148
K R W B E = 158
K R L D E = 156
KR L B E = 155
K R H D E = 182
K R H B E = 175
K S W D E = 141
K S W B E = 151
K S L D E = 140
K S L B E = 139
K S H D E = 144
K S H B E = 137
Therefore, the preferred route to minimise the distance travelled is the route K S H B E.
Now, we come back to our case. What we want is the minimised miles, within an expected time spending on the road. The function could be written as:
L stands for the total driving distance;
stands for the distance between dustbin (N-1) to dustbin N;
T stands for the total time;
stands for the driving time from dustbin (N-1) to dustbin N;
t stands for the maximised expected time spent on the collection for each time;
N stands for all bins being considered, which means their ER are over 80%.
By programming such a DP code, we can get the result. This result could be used for a guidance for waste collection.
-GIS Shortest Path approach
As we discussed before, we can also use the SPP to analyse this situation and get a corresponded result.
The SPP has been broadly used by many companies and apps, for instance, Mapquest and Google Maps. There is also a software being developed, called ArcGIS. If we could install this app on each truck, and negotiate with the software owner to make slight changes that will better suit the waste collection case, there will be a bright future for developing the smart waste collection in the city.
However, there are also two main challenges. The first one is how to maintain such a big real time database. As it is common for a city to have millions of dustbins, even if just for a specific area, the amount of bins are also too large to manage. Also, what frequency should we keep to update the real time data? For how long should we keep the historical data? To be more practical, should we separate the bin for glass, plastic, bottles and bio-residuals? What should we do if there is heavy density inside the bin that might not be able to fully reflect the volume? Some of these questions can be answered by the accumulation of first-hand experiences, some of the answers might be more tricky and more difficult to find out.
The second challenge is that which organisation should be the regulator? Should it be city or different from city ?
Besides these two main challenges, there is also other technique problems, as how to update and upgrade the code and how to cooperate with the IT solution provider developing app. These could be the great opportunity, while also have the chance to be the obstacle.